column chromatography. .r\ portion of solution E was lyophilized. and 100 mg. of the dried substance was placed on a silicic acid column (1 X 9 cm.). The column \vas developed iiith benzeneethyl ether-ethyl alcohol. 50:15:2. The eluted fractions were collected and xveighed. Progress in rhe fractionation was followed by means of glass fiber paper chromatograms, using the equipment and procedures described by Dieckert and Reiser ( 3 ) and modified by Dieckert and Morris (1). The glass fiber p a l m (Reeve Angel. S o . x-934-.4H) \cas impregnated Siith silicic acid ( I ) . T h e developing solvent for the paper chromatograms was the same as that used to develop the silicic column. &After the solveni had evaporated from rhe papcr. the chromarogram !vas sprayed lvith concentrated sulfuric acid and heated. first mildly, over a hot plate for thc development of color, and finally more vigorously to char the organic substaiices to produce black spots on the ivhite background (5). .\bout the first 8'36 of the dried eluate from the column contained substances that moved with the solvent front on the glass paper. T h e next. 15Yc shoired a spot \vel1 separated from both the front and the origin. This spot showed a pink to violet color Lrhen the chromatogram \vas sprayed Lrith concentrated sulfuric acid and heated mildly. The Ri for the spot corresponded to that produced by an authentic specimen of $sirosterol-D-glucoside. In Figure 2,
A is a chromatogram of solution E, B @-sitosterol-D-glucoside shoM ed the I\\ o one of authentic ~-sitosterol-~-g~ucoside, compounds to be identical (Figure 3). The infrared spectra of the aglycone of C one of the glucoside separated from the silicic acid column, and D one of the glucoside from peanut flour and of the glucoside cr) s t a l k e d from solution authentic @-sitosterolshowed these compounds to be identical also (Figure 4). E. The glucoside was hydrolyzed according to the method of Power and Acknowledgment Salway (6). Five milligrams i t as added to a mixture composed of 3 ml. of amyl The authors express their appreciaalcohol, 5 ml. of 157, aqueous HCI. tion to L. J. S\iift for an aurhenric and 25 ml. of methanol. The resulting sample of B-sitosterol-n-glucoside ; to solution !vas refluxed for 6 hours and the L. E. Brown. Elizabeth R . McCall. amyl alcohol remo\red by steam distillaand Donald Mitcham for chemical and tion. ph>-sical analyses; and ro .I.J. C:rovetto The aglucone ivas removed from the for defatting the peanuts. aqueous solution by extraction Lvith diethyl ether. I t was then crystallized Literuture Cited from methanol (m.p. 136-38" C . ) , with no depression of the melting point (1) Dieckert. J. IV.. Carney. \V, B.. on admixture \iith authentic 8-sitosterol. Ory. R. L.. Morris. S.J . . .-lnnl. ChPrn. Analysis: C, 83.0; H: 12.3. Calculated 30, 1442 (1958). (2) Dieckert. J. b'..hforris, S..J.. Ihid., for @-sitosterol: C. 83.99; H, 12.15. 29, 31 (1957). The aqueous solution was treated (3) Dieckert, J. IV.% Iforris. S. J.. J. \vith an excess of silver carbonate and AGR. FOODCHEM.6, 930 (19.58). filtered. The filtrate was chromato(4) Dieckert, J. \\-.. Reiser. R.. J . .Am. graphed on glass fiber paper rrhich had Oil Chemists' Soc. 33, 123 11956). been impregnared with monopotassium (5) Hamilton, J. G.. Dieckert. .T. \V., phosphate in accordance Ivith the Arch. Biochem. Biopl7~s.82, 203 (19.39). method of Dieckert and Morris (,?). (6) Po\rer, F. B.. Saljvay. .A. H.. J . The R, value for the sugar obtained from CI7~rn.5'06. 103, 399 (19131. the hydrolyzate \vas the same (0.78) Received for relieu: hhuember 7. 1900. A c as that obtained with D-glucose tvhen the cupted January 23. 1967. 76th Southriurchromatogram was developed Ivith aceRegional &Meeting.ACS, Oklahoma City. O k l a . , December 7 9 6 0 . T h e mention of namrs o j tone-n-propyl alcohol-iiater, 50:40 :I 0. firms or trade products does not impiy / h a / thry The infrared spectra (potassium broare endorsed or recommended by the D ~ p a r ~ r n r n t mide disk method) of the glucoside isoof .4gricuiture oiler o t h u j r m s or riniilai pron'ucls lated from peanut flour and of authentic not nientionfd.
CAROTENOID EPOXIDE DETECTION
I An Improved Test for Carotenoid Epoxides
W
HES ETHER solutions of certain carotenoids are shaken with concentrated hydrochloric acid, the acid layer turns blue (7). Most of these carotenoids are epoxides (a number of \vhich occur in nature) or their furanoid transformation products, which are referred to here as 5,6-epoxides and 5,8epoxides. respectively, following the nomenclature of Goodliin (,5). In genvral. a much deeper blue color forms \iith diepoxides than with monoepoxides ( 7 ); hydroxyl groups increase the stability and intensity of the blue color ( 6 ) . An improved test has been developed for detection of carorenoid epoxides and. for differentiation of diepoxides and monoepoxides; a modification of this
I
test can be used to distinguish 5.6epoxides from 5,8-epoxides. Experimental
Materials. The carotenoid fractions used \$ere obtained from known sources. especially oranges ( 4 ) . cling peaches ( 7 ) , and leaves ( 3 ) ; capsanthin and capsorubin were obtained from red garden peppers, and cryptoxanthin mono- and diepoxides from Meyer lemons ( 2 ) . The identifications were based on spectral absorption curves, and on behavior in countercurrent distribution. chromatography. and acid treatment. A sample of ,5'-apo-2-carotenal was kindly provided by the late G. F. Siemers, Hoffman-La Roche Inc., Nutlel-. S. J.
A. LAURENCE CURL and GLEN Western F. Regional Research
Observations on the Hydrochloric Acid-Ether Test. Hydrochloric acidether tests were made by the procedure previously described ( I ) on many carotenoid fractions obtained from various fruits and leaves. .4 deep blue color was formed by violaxanthin (zeaxanthin 5,6,5 ',6 '-diepoxide), its 5.8-epoxide isomerization products. and some poly01 monoepoxides. In other cases (such as cryptoxanthin diepoxides and neoxanthin) a definite, but jveaker color was formed. and in still other cases (mutatoxanthins and cryptoxanthin monoepoxides) the blue color \vas pale or uncertain. Capsanthin (a diol kerone) gave a light salmon color in the liydrochloric acid layer, capsorubin (a diol
V O L . 9, NO. 5, S E P T . - O C T .
1961
403
An improved hydrochloric acid color test for carotenoid 5,6- and 5,a-epoxides has been developed in which the reaction is carried out in one phase of ether-methanol. Diepoxides give greenish blue colors; monoepoxides, usually yellow-green; while two monoepoxides of unknown structure, persicaxanthin and sinensiaxanthin, give red-violet colors. A modified test, using citric acid instead of hydrochloric acid, can be used to detect the presence of 5,6-epoxides. Both of these tests can be readily followed spectrophotometrically. Contrary to earlier reports, neither P-apo-2-carotenal nor capsorubin gave a blue color in the hydrochloric acid-ether test.
diketonel a stronger salmon color, and trollein (a nonepoxide polyol) ( 4 ) a rather strong yellow color; some polyol monoepoxides gave green colors. 0Apo-7-carotenal gave a negative test. I t was reported that the latter substance gi\.es an intense, stable blue coloration in this test (8); while capsorubin which was reported earlier (7, 9 ) yields a deep blue color. It appears that for the last tbvo. the preparations used in the earlier work had epoxide impurities. Ether is much more soluble in concentrated hydrochloric acid than it is in water; about 2.2 volumes of ether (L*.S.P.l dissolve in one volume of acid. The resulting solution has an appreciable solubility for polyoxygen carotenoids ; hence these may give a color in the loiver layer: although they are not epoxides. The substance or substances responsible for the blue color are unknown. A solution of violaxanthin in ether was extracted repeatedly Tvith concentrated hydrochloric acid until the extract was only a pale blue. The extracts were added to a mixture of ether and water, and potassium hydroxide in methanol lvas added until the blue color disappeared: on mixing, the ether layer became red-orange. The recovered pigment on chromatography yielded a number of bands. The strongest appeared to be a cis-mutatoxanthin (spectral absorption maxima in benzene, 460. 434, and 319 m p ) , but the others did not resemble known carotenoids. Hence! the blue compound did not revert to the 5,8,5 '?8'-diepoxide on neutralization. The Hydrochloric Acid EtherMethanol Test for Carotenoid Epoxides. Formation of a blue color with h!-drochloric acid and carotenoid epoxides occurs in other solvents such as methanol. ethanol, and acetone. It forms rapidly in acetone: but with some Carotenoids there is a considerable increase in yellow or orange color. Not all carotenoids are readily soluble in methanol or ethanol; a 1 : l mixture of ether and methanol will dissolve both carotenoids and concentrated hydrochloric acid. The color formed is not affected by partition of the carotenoid betiveen the two phases as in hydrochloric acid-ether. A convenient amount of
-
404
AGRICULTURAL AND
0
, 550
VA I
600
650
WAVE L E N G T H ,
rnp
700
750
Figure 1. Absorbance a t wave lengths 500 to 750 mp of solutions of carotenoids in ethyl ether-methanol-concentrated hydrochloric acid, 10 to 9 to 1 b y volume N. Neoxanthin, 7 hours
VA. Valenciaxanthin, 4 hours 35
minutes P. Persicaxanthin, 40 minutes VI. Violaxanthin, 7 hours S. Sinensiaxanthin, 2 hours 20 minutes Initial absorbance of controls a t principal maxima ca. 1.9
carotenoid is that equivalent to 20 ml. of solution having an absorbance a t the principal maximum of 1 in a 1-cm. cell. The procedure is as follows: Evaporate two equal aliquots of carotenoid solution in vacuo, preferably in a rotary evaporator. Dissolve the residues in 10 ml. of ether (preferably U.S.P., peroxide free), and add 9 ml. of methanol. Add 1 ml. of water to one solution as a control, and 1 ml. of concentrated hydrochloric acid to the other. Compare the colors of the two solutions in similar test tubes. The acid solution will often become greener at once if an epoxide is present. At the end of 1 hour, acid-treated diepoxide solutions will be greenish blue, monoepoxides usually yellowish green. Nonepoxides are usual!y a little paler than the controls, but some cis-isomers may sholv an increase in the yellow or orange color, as does also valenciaxanthin (a 5,6-epoxide). A red-violet color indicates presence of epoxides such as persicaxanthin or sinensiasanthin with FOOD CHEMISTRY
a shorter conjugated double bond system. I n the above test? the blue colors formed \zith typical mono- and diepoxides are similar, but the residual color of the carotenoid is different. 5,6-Monoepoxides on treatment with acid are rapidly converted to 5,8-eposides with a decrease in wave length of the spectral absorption maxima of about 20 mp, while with 5,6,5 ',6 '-diepoxides the decrease is over twice as great. The color change is more pronounced with 5>6-epoxidesthan with 5.8-epoxides. Test for Carotenoid 5,6-Epoxides. T h e follo\zing simple test for distinguishing 5,6-epoxides from 5,g-epoxides or nonepoxides is based on the shift in spectral absorption maxima of the 5,6-epoxides on contact Lvith acid. I n order to avoid formation of a blue or green color, citric acid is used instead of hydrochloric acid. The procedure is as follows : Pipet out equal aliquots of the carotenoid solution (a convenient amount is the equivalent of 10 ml. of solution with
a n absorbance ai. the principal maximum of 1 in a 1-cm. cell). Evaporate in vacuo and dissolve the residues in 5 ml. of ether and transfer to test tubes of the same diameter. To one, add 5 ml. of methanol and to the other, 5 mi. of a 1% solution of citric acid in methanol. Stopper and allow to stand for 1 hour. If a 5:G-epoxide is present. the solution with acid added will become definitely paler when compared cross\vise. Slight decreases in color may be observed in nonepoxide carotenoids \Then viewed vertically. probably due to trans-cisisomerization. IVith valenciaxanthin and persicaxanthin, which have very pale colors initially, there is a marked increase in greenish fluorescence in ultraviolet light. This test is positive for 5;6-. 5.6.5l.6 ’-, and 5,6?3‘:8 ’-epoxides, but negative for 5,8- and 5$,5 ’:8’-epoxides and for nonepoxides. In this test. 5,6.5‘:6’diepoxides show a much greater color change than 5,6-epoxides or 5.6,s ’:8’diepoxides. Spectrophotometric Observations. T h e above tests can conveniently be adapted to a spectrophotometer. T h e development of the blue color was studied for numerous carotenoid epoxides, using a Cary Model 14. T h e absorbance at the principal maximum of the control (10 ml. of U.S.P. ether, 9 ml. of methanol, and 1 ml. of lvater) was about 1.9 in a 1-cm. cell. A duplicate aliquot was evaporated in vacuo, the residue dissolved in 10 ml. of ether and 9 ml. of methanol, and 1 ml. of concentrated hydrochloric acid was added. Spectrophotometric curves were run as soon as possible, and then at intervals u p to 24 hours, over the range 750 to 300 mp. I n the initial curves, the 5,6-epoxides showed a complete conversion to 5,8-epoxides and a single broad maximum appeared a t lvave lengths above 550 m p (Figure 1 ) . T h e latter increased in magnitude with time, in some cases reaching a value 50 o r 60% of that of the principal maximum of the control in 24 hours or less, while the maxima of the 5,8-epoxide decreased a n d eventually disappeared. I n most cases, the blue color was still present after several days. I n Table I are given comparative data on the effect of hydrochloric acid in methanol-ether on 14 carotenoids. T h e positions of the long wave length maxima Irere at wave lengths of 6’0 to 676 m p
Table 1.
Absorption Maxima of Blue Colors Formed by Addition of Hydrochloric Acid to Solutions of Carotenoids in Methanol-Ether Absorption M a x i m a Formed in Presence of Hvdrochloric Acid absorbance Wave o f principal Time, length, maximum of hr. Mu control
yo
No.
1
2 3 4 5 6
I
8 9 10 11 12
Component
Type o f Epoxide
Absorbance M a x i m a of Control
Cryptoxanthin Cryptoxanthin Mutatoxanthin “ Mutatoxanthin Violaxanthin Luteoxanthin Luteoxanthin Xuroxanthins Seoxanthin Neochrome Neochrome Valenciaxanthin
5,8-epoxide 5,8,5 ’,8 ’ diep. (5. 8) (5. 8) ( 5 . 6. 5‘. 6’) (5. 6. 5 ’ . 8’) (5. 6 . 5‘. 8’) ( 5 . 8. 5’. 8’) ( 5 , 6) ( 5 . 8) (5. 8 ) (5. 6)
451, 426, (404)a 426, 401 . 375 450, 424, (404) 451, 426, (406) 470, 439, 416 449, 422. 399 448, 422, 398 426. 401, 380 465. 436. 413 448. 422, 398 448. 421, 398 390, 369, 351
1
1 1 1
0.25 1
1 1 1
1 0.25 1
673 623 670 676 667 654 6’4 636 665 666 667
15
23 5 12 13 26 16 26 10 7
17 1 .5 (at 545 m u ) 566 29 586 36
392, 3’1, 353 1 (5. 6) 0.67 427, 402. 3805 ( 5 . 6) n Values in parenthesis represent shelves or humps (not maxima) on spectral absorption curves. h In benzene. 13 Persicaxanthin 14 Sinensiaxanthin
for derivatives of @-carotene monoepoxide (numbers 1, 3. and 4) and a t 665 to 667 for those of a-carotene monoepoxide (numbers 9, 10. and 11). Positions of the blue maxima were much more variable for derivatives of @-carotene diepoxide (numbers 2. 5. 6, 7, and 8), ranging from 623 to 674 mp. I n Table I. data are included for three isomeric 5,8-epoxide pairs, numbers 3 and 4, 6 and 7, and 10 and 11. I n all three cases, the blue color formed in one of each pair a t about twice or more the rate of the other (right-hand column). T h e positions of the long wave length maxima with sinensiaxanthin and persicaxanthin (two poly01 5,6-epoxides of incompletely known structure obtained from oranges and cling peaches, respectively) were a t much shorter wave lengths. These maxima developed and then faded much more rapidly than in numbers 1 to 11, L’alenciaxanthin was unique in that after acid treatment there was a small amount of absorbance from 500 to 750 m p with no maximum (Figure 1). I n the hydrochloric acid-ether test it formed a light blue color (4). Persicaxanthin and valenciaxanthin can hence be differentiated by treatment with hydrochloric acid as uell as by
countercurrent distribution 7). The valenciaxanthin solution in hydrochloric acid-erher-methanol became more yellow on sranding, Lvith maxima appearing a t 416 and 390 mp. \\.ith an inflection at around 466 mp. 1
Literature Cited
(1) Curl, A. L.: Food Research 24, 113 (1 959). (2) Curl, .4. L.? unpublished ~ t - o r k , M’estern Regional Research Laboratory. Albany, Calif., 1958. (3) Curl, A. L., Bailey: G. F.! Food Research 22, 323 (1957). (4) Curl, A. L., Bailey, G . F., J. . ~ G R . FOODCHEM.2, 685 (1954). (5) Goodwin, T. W., “Modern Sferhods of Plant Analysis,” Vol. 3, p. 2-1, Springer L’erlag, Berlin? 1953. (6) Karrer, P., Fortschr. Chrin. 9’4. .Yaturstnfe 5, 1 (1948). (7) Karrer, P., Jucker. E.. ..Ca:-otenoids:” p. 7, Elsevier, S e i s York. 1950. (8) Karrer, P., Solmssen, Y . ? Hfh. Chim. A c t a 20, 682 (1937). (9) Zechmeister, L.. Cholnoky. L. v., A n n . Chem.-Liebigs 516, 30 (1935). Receiwd f o r review August 26, 79ri0. .4rcepted February 70, 1967. TIorX doiw ot ‘I laboratory of the Agricultural Rrseaich .Siri,ice, c‘. 5. Department of Agricultur~. l i r gf trade names of specific materials GT q u i p i ~ i z ! does not constitute a recommendation b y :he U. S. Department of Agricultur?.
V O L . 9, NO. 5
SEPT.-OCT.
1961
405
